Pharmacological advances in multi-targeted strategies for type 2 diabetes mellitus: a systematic perspective based on traditional Chinese medicine.

Type 2 diabetes mellitus (T2DM) is a complex systemic metabolic disease driven by insulin resistance, β-cell dysfunction, chronic low-grade inflammation, oxidative stress, and neuro-immune dysregulation. It frequently progresses to multi-organ complications affecting the kidneys, retina, heart, and central nervous system. This review synthesizes mechanistic and translational evidence on Traditional Chinese Medicine (TCM)-related botanical drugs and botanical preparations (formula-based interventions), along with representative plant metabolites that are frequently investigated in the TCM research context (e.g., berberine, baicalin, and tanshinone IIA, which are not unique to TCM). For formula-based preparations, we extracted and reported intervention identity elements (dosage form, complete composition, and processing/standardization as described in primary studies); missing identity items were recorded as not reported (NR) and not inferred. We organized findings across shared T2DM-relevant pathogenic modules, including PI3K/Akt and AMPK signaling, inflammatory outputs (NF-κB/NLRP3), redox regulation (NRF2/ROS), angiogenic signaling (VEGF), and gut-liver-brain-immune network interactions, emphasizing studies in which pathway modulation is accompanied by metabolic or complication-relevant endpoints. To strengthen interpretability and reproducibility, we conducted a structured literature search (2000-2025) and applied evidence grading (human/RCT vs. animal vs. in vitro/in silico), and we critically appraised reporting quality using the GA-online Best Practice in Research - ConPhyMP tool. All source organisms were taxonomically validated using authoritative resources, and full scientific names (including author citation and family) were standardized. We caution that compound-target links, particularly those derived from in silico predictions or single-assay readouts, may be vulnerable to assay interference liabilities (including PAINS) and should be supported by orthogonal validation and outcome-linked readouts before strong mechanistic claims are made. Finally, we outline translational priorities, including rigorous standardization and quality control (distinguishing analytical marker metabolites from bioactive metabolites), improved study design and controls, and well-designed randomized, pragmatic, and real-world evaluations with clinically meaningful endpoints (e.g., HbA1c, complication progression, and safety).